The reliability of power in physical performance tests is crucial for accurately assessing athletes, patients, and others. This review identifies the most reliable measures of power and the factors affecting reliability. The study analyzed 101 studies on healthy adults, finding that tests like sprint running, peak power on a treadmill or cycle ergometer, and lactate-threshold power had the smallest coefficient of variation (CV), around 0.9%. However, CV was higher for nonathletes, females, shorter or longer tests, and respiratory-based measures. Practice trials significantly improved performance, with a 1.2% improvement between the first two trials, but only 0.2% between subsequent trials. High reliability is essential for accurate performance assessment, as unreliable tests cannot track changes or provide precise single-trial results. The review emphasizes the importance of using reliable tests for accurate measurement. Reliability refers to the consistency of test results when repeated. The study reviewed laboratory and field tests of physical power, focusing on athletes but including all studies on healthy adults. Different measures of reliability were converted to two main metrics: standard error of measurement and mean performance change between trials. Validity, which relates to whether test results reflect true performance, was not addressed due to the need for more theoretical work. However, reliable tests are necessary for valid results. The methods used to assess reliability are based on recent reviews, with the main measure being the typical percent error, or standard error of measurement.The reliability of power in physical performance tests is crucial for accurately assessing athletes, patients, and others. This review identifies the most reliable measures of power and the factors affecting reliability. The study analyzed 101 studies on healthy adults, finding that tests like sprint running, peak power on a treadmill or cycle ergometer, and lactate-threshold power had the smallest coefficient of variation (CV), around 0.9%. However, CV was higher for nonathletes, females, shorter or longer tests, and respiratory-based measures. Practice trials significantly improved performance, with a 1.2% improvement between the first two trials, but only 0.2% between subsequent trials. High reliability is essential for accurate performance assessment, as unreliable tests cannot track changes or provide precise single-trial results. The review emphasizes the importance of using reliable tests for accurate measurement. Reliability refers to the consistency of test results when repeated. The study reviewed laboratory and field tests of physical power, focusing on athletes but including all studies on healthy adults. Different measures of reliability were converted to two main metrics: standard error of measurement and mean performance change between trials. Validity, which relates to whether test results reflect true performance, was not addressed due to the need for more theoretical work. However, reliable tests are necessary for valid results. The methods used to assess reliability are based on recent reviews, with the main measure being the typical percent error, or standard error of measurement.